Following the closure of google fusion tables this December, I want to know if there are any alternative to it working with data collected in Collect Earth. In particular, how to train classifiers without Fusion Tables. asked 16 Dec '19, 09:13 BrightEdem |
Google Fusion Tables are deprecated and as you say cannot be used anymore in GEE. Luckily the data collected with Collect Earth can still be used in GEE by uploading it as a table. When you go in GEE Code Editor you can see that in the tab Assets you have the possibility to upload the data as a CSV (Table ) : When you do that, there is a dialog where you have to specify that the columns for the X and Y coordinates are XCoordinate and YCoordinate After you upload the table you can reference it in the ee.FeatureCollection using the Asset id ( where you do not need anymore the ft: prefix as that was used to indicate the data came from a Fusion Table) Let us know if it there are issues using the data! answered 16 Dec '19, 10:45 Open Foris ♦♦ Thanks for the feedback. How do i reference the Asset ID in the ee.FeatureCollection?. Will this enable me to train my classifiers?
(16 Dec '19, 14:02)
BrightEdem
|
This is a link explaining how to work with Assets : https://developers.google.com/earth-engine/asset_manager This is another link to a Google Earth Engine script which uses cart to classify Lansat 8 imagery. You can test it with your data by changing the line 31 ( use the Asset you have uploaded) and also line 37 where you specify what property from each feature ( or column for each line in the CSV) should be used to classify with! Here is another link that might also be useful from the Google Earth Engine tutorial. For more info about the classification algorithms it is better to ask in the GEE discussion group ( search in the archive before asking ). answered 16 Dec '19, 14:52 Open Foris ♦♦ I followed the script and suggestions you put forward, thus i change line 31 with my asset id, line 37 with the column name in my csv i want to use. I however get the errors "classification: Layer error: No data was found in classifier training input", "Dictionary (Error) No data was found in classifier training input"
(19 Dec '19, 09:12)
BrightEdem
|